The Genetic Lottery Goes To School: Better Schools Compensate for the Effects of Students’ Genetic Differences. The data is getting better. Schools help the students at the bottom improve a bit in reading but not numeracy.
Superhuman Performance of a Large Language Model on the Reasoning Tasks of a Physician. They said this day was coming soon, and it's here. LLM Diagnostics perform much better than actual physicians.
Black-White IQ gap sure looks genetic.
Bicycle Helmets. Ummm.
"At triage, nurse-recorded data included sex, age, chief complaint, presumptive diagnosis, triage nurse note, acuity number, means of arrival, and initial vitals. At the end of the emergency department
ReplyDeletevisit, the provider's history of present illness, physical exam, medical decision making, imaging reports, and labs were collected. In many cases the provider was a resident, in which case theattending attestation history of present illness, physical exam, and medical decision making was
also captured. Once the patient arrived on the medical floor or medical ICU, the admission chief complaint, history of present illness, physical exam, and assessment and plan were captured. Similarly to the emergency room, many admission notes were completed by residents in which
case the attending attestation history of present illness, physical exam, and assessment and plan were also captured"
So sounds like human skill and judgment were indeed involved in captuiring this data...also the diagnosis info from previous encounters.
In the bicycle helmets article, they mention "Claims of risk compensation relating to mask-wearing at the start of the COVID-19 pandemic" but I don't see anything relating to risk-compensation in the (UK) Times article they reference, just to risks (poor mask hygiene).
ReplyDeleteThe only 'wayback-machine' saved version is from April of this year, so it may be that the article was stealth-edited after the original "the studies show" reference to it.
A friend who was a pharmacist (worked in a hospital as well as retail pharmacy) remarked that in her experience, doctors didn't know much about drugs. I think maybe LLMs, perhaps coupled with individualized simulation models ('digital twins') might improve the matching of drugs to conditions, especially when the same patient has a lot of medical issues.
ReplyDelete